CN114693684A - Airborne fan blade defect detection method - Google Patents

Airborne fan blade defect detection method Download PDF

Info

Publication number
CN114693684A
CN114693684A CN202210613508.3A CN202210613508A CN114693684A CN 114693684 A CN114693684 A CN 114693684A CN 202210613508 A CN202210613508 A CN 202210613508A CN 114693684 A CN114693684 A CN 114693684A
Authority
CN
China
Prior art keywords
hoyer
fan blade
sliding window
current
pixel point
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202210613508.3A
Other languages
Chinese (zh)
Inventor
李小刚
李峰平
林苏奔
孙浩然
张昆鹏
郭剑
邵正鹏
李函禧
卢成绩
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Liaiwei Innovative Intelligent System Zhejiang Co ltd
Original Assignee
Liaiwei Innovative Intelligent System Zhejiang Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Liaiwei Innovative Intelligent System Zhejiang Co ltd filed Critical Liaiwei Innovative Intelligent System Zhejiang Co ltd
Priority to CN202210613508.3A priority Critical patent/CN114693684A/en
Publication of CN114693684A publication Critical patent/CN114693684A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0004—Industrial image inspection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/12—Edge-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/13—Edge detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/136—Segmentation; Edge detection involving thresholding
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30108—Industrial image inspection

Abstract

The invention discloses a defect detection method for an airborne fan blade, which comprises the following steps of firstly obtaining a surface image of the fan blade, then converting the surface image into a gray image, then continuously traversing all pixels of a picture to generate a Tenengrad gradient matrix, judging a fan blade area based on a Tenengrad gradient threshold value, then traversing all pixels of the current picture to generate a Hoyer statistical value matrix, further reserving Hoyer statistical values of the fan blade area in the step according to the judgment condition of the fan blade area, setting the Hoyer statistical values of other areas to zero, finally setting a Hoyer statistical value threshold value, and if the Hoyer statistical value is larger than the threshold value, determining that the pixel point is a defect characteristic. The method directly identifies the defects of the fan blade from the Tenengrad gradient and the Hoyer statistical value, and has the characteristic of being not influenced by the distribution of training data samples.

Description

Airborne fan blade defect detection method
Technical Field
The invention relates to the technical field of surface quality detection of wind power facilities, in particular to a method for detecting defects of blades of an airborne fan.
Background
The fan blade is used as a wind catching mechanism of the wind driven generator and can be influenced by the fan blade or the outside in the operation process. If the wind power blade is damaged, the wind power blade is light and becomes economic loss, and is heavy and generates potential safety hazard. In the conventional monitoring method of the fan blade at present during regular inspection, maintenance personnel use a telescope to observe the state of the blade at a long distance and record the state on a log. However, the manual detection method inevitably has the problems of long detection time, low precision, high cost and the like, and is difficult to detect the large-scale fan blades of the wind field.
Disclosure of Invention
The invention aims to provide a method for detecting defects of an airborne fan blade. The invention can realize reliable and efficient state monitoring and defect identification of the fan blade and has the characteristic of no influence of training data sample distribution.
The technical scheme of the invention is as follows: a method for detecting defects of an airborne fan blade comprises the following steps:
s1: acquiring a surface image of a fan blade, and converting the surface image of the fan blade into a gray image;
s2: setting a sliding window with the window size of w1 multiplied by w1, traversing all pixels of the current gray picture by using the sliding window, and calculating the Tenengrad gradient of the pixels in the current window to form a Tenengrad gradient matrix;
s3: setting a Tenengrad gradient threshold, and if the Tenengrad gradient is greater than the threshold, setting the pixel point as a fan blade area;
s4: setting a sliding window with the window size of w2 multiplied by w2, traversing all pixels of the current gray picture by using the sliding window, and calculating the Hoyer statistical value of the pixels in the current window until all pixels of the current picture are traversed to form a Hoyer statistical value matrix;
s5: keeping the Hoyer statistical value of the fan blade area in the step S4 and setting the Hoyer statistical values of other areas to zero;
s6: and setting a Hoyer statistical value threshold, and if the Hoyer statistical value of the fan blade area is greater than the threshold, determining that the pixel point is a defect characteristic.
In the method for detecting the defects of the blades of the airborne fan, the algorithm of the Tenengrad gradient is as follows:
Figure DEST_PATH_IMAGE001
wherein
Figure 584691DEST_PATH_IMAGE002
As the Tenengrad gradient information for the current sliding window,
Figure DEST_PATH_IMAGE003
，
Figure 285931DEST_PATH_IMAGE004
for the row and column index numbers of the sliding window of the image respectively,
Figure DEST_PATH_IMAGE005
is the size of the window or windows,
Figure 244529DEST_PATH_IMAGE006
is a point
Figure DEST_PATH_IMAGE007
The information of the gradient in the x-direction,
Figure 132850DEST_PATH_IMAGE008
is a point
Figure 100002_DEST_PATH_IMAGE009
Gradient information along the y-direction.
In the method for detecting the defects of the blades of the airborne fan, the judgment algorithm of the fan blade area is as follows:
Figure 467885DEST_PATH_IMAGE010
；
wherein
Figure DEST_PATH_IMAGE011
The judgment result of the current pixel point is that 1 represents that the current pixel point is positioned on the fan blade, 0 represents that the current pixel point is positioned outside the fan blade area,
Figure 758053DEST_PATH_IMAGE003
，
Figure 435022DEST_PATH_IMAGE004
for the row and column index numbers of the sliding window of the image respectively,
Figure 331345DEST_PATH_IMAGE012
the Tenengrad gradient corresponding to the current pixel point,
Figure DEST_PATH_IMAGE013
is a preset threshold of Tenengrad gradient.
In the method for detecting the defects of the blades of the airborne fan, the algorithm of the Hoyer statistical value is as follows:
Figure 955225DEST_PATH_IMAGE014
wherein
Figure DEST_PATH_IMAGE015
For the Hoyer statistic for the current sliding window,
Figure 614745DEST_PATH_IMAGE003
，
Figure 462615DEST_PATH_IMAGE004
for the row and column index numbers of the sliding window of the image respectively,
Figure 856688DEST_PATH_IMAGE016
is the size of the window(s),
Figure 549837DEST_PATH_IMAGE017
is a point
Figure 100002_DEST_PATH_IMAGE018
The gray scale information of (2).
In the method for detecting the defects of the blades of the airborne fan, in step S5, the algorithm for retaining the Hoyer statistical value is as follows:
Figure 532705DEST_PATH_IMAGE019
wherein
Figure DEST_PATH_IMAGE020
For the kept Hoyer statistics,
Figure 20319DEST_PATH_IMAGE021
is a matrix of the Hoyer statistical values,
Figure DEST_PATH_IMAGE022
is the judgment result of the pixel points in the fan blade area,
Figure 619796DEST_PATH_IMAGE003
，
Figure 116637DEST_PATH_IMAGE004
are the row-column index numbers of the image sliding window respectively.
In the method for detecting the defect of the onboard fan blade, the algorithm for judging the defect characteristics of the fan blade is as follows:
Figure 32640DEST_PATH_IMAGE023
；
wherein the content of the first and second substances,
Figure DEST_PATH_IMAGE024
is the defect judgment result of the current pixel point, 1 represents that the current pixel point is the defect position, 0 represents that the current pixel point does not contain defect information,
Figure 409263DEST_PATH_IMAGE003
，
Figure 777928DEST_PATH_IMAGE004
for the row and column index numbers of the sliding window of the image respectively,
Figure 812880DEST_PATH_IMAGE025
the Hoyer statistic value after being reserved for the current pixel point,
Figure DEST_PATH_IMAGE026
is a preset Hoyer statistical value threshold value.
According to the method for detecting the defects of the blades of the airborne fan, the sliding window is moved to the upper left corner of the gray image, and one pixel is sequentially moved from left to right and from top to bottom until the sliding window traverses all pixels of the current gray image.
Compared with the prior art, the invention has the following beneficial effects:
the method comprises the steps of firstly obtaining a fan blade surface image, secondly converting the fan blade surface image into a gray level image, then traversing all pixels of a picture by using a sliding window to generate a Tenengrad gradient matrix, judging a fan blade area based on a Tenengrad gradient threshold, then continuously traversing all pixels of the current picture by using the sliding window to generate a Hoyer statistical value matrix, further reserving Hoyer statistical values of the fan blade area in the step according to the judgment condition of the fan blade area, setting the Hoyer statistical values of other areas to zero, finally setting a Hoyer statistical value threshold, and if the Hoyer statistical value is larger than the threshold, determining that a pixel point is a defect characteristic. The method directly identifies the defects of the fan blade from the Tenengrad gradient and the Hoyer statistical value, and has the characteristic of being not influenced by the distribution of training data samples. The invention can realize reliable and efficient state monitoring and defect identification of the fan blade.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a fan blade surface image used in example 1 of the present invention;
FIG. 3 is a Tenengrad gradient matrix formed in example 1 of the present invention;
fig. 4 is a fan blade area judgment result obtained in embodiment 1 of the present invention;
FIG. 5 is a Hoyer statistic matrix formed in example 1 of the present invention;
FIG. 6 shows the results of judging the defective characteristics in example 1 of the present invention;
FIG. 7 is a fan blade surface image used in example 2 of the present invention;
FIG. 8 is a Tenengrad gradient matrix formed in example 2 of the present invention;
fig. 9 is a fan blade area judgment result obtained in embodiment 2 of the present invention;
FIG. 10 is a Hoyer statistical value matrix formed in example 2 of the present invention;
fig. 11 shows the judgment result of the defect characteristics in example 2 of the present invention.
Detailed Description
The invention is further illustrated by the following figures and examples, which are not to be construed as limiting the invention.
Example 1: the invention relates to a method for detecting defects of an airborne fan blade, which is further described by combining specific cases, and a flow chart of the method is shown as a figure 1, and comprises the following steps,
1) fixing a shooting device on the unmanned aerial vehicle, and shooting a fan blade to obtain a surface image of the fan blade, as shown in the attached figure 2;
2) converting the blade surface image to a grayscale image;
3) setting the sliding window width to be w1, forming the sliding window with the window size of w1 xw 1, wherein the window width is 21 in the example;
4) moving the sliding window to the upper left corner of the gray image, and calculating the Tenengrad gradient of the elements in the current window (the Tenengrad gradient function adopts a Sobel operator to respectively extract gradient values in the horizontal direction and the vertical direction), wherein the algorithm of the Tenengrad gradient is as follows:
Figure 317811DEST_PATH_IMAGE027
wherein
Figure DEST_PATH_IMAGE028
As the Tenengrad gradient information for the current sliding window,
Figure 865336DEST_PATH_IMAGE003
，
Figure 721296DEST_PATH_IMAGE004
for the row and column index numbers of the sliding window of the image respectively,
Figure 825518DEST_PATH_IMAGE005
is the size of the window or windows,
Figure 434223DEST_PATH_IMAGE006
is a point
Figure 700119DEST_PATH_IMAGE007
The information of the gradient in the x-direction,
Figure 777797DEST_PATH_IMAGE029
is a point
Figure DEST_PATH_IMAGE030
Gradient information along the y-direction.
5) Moving the sliding window to the right by one pixel, and calculating the Tenengrad gradient of the element in the current window;
6) continuously traversing all pixels of the current picture to form a Tenengrad gradient matrix, as shown in figure 3;
7) setting a Tenengrad gradient threshold, if the Tenengrad gradient is greater than the threshold, taking the pixel as a fan blade area, and adopting a fan blade area judgment algorithm as follows:
Figure 154552DEST_PATH_IMAGE031
；
wherein
Figure 883342DEST_PATH_IMAGE011
The judgment result of the current pixel point is that 1 represents that the current pixel point is positioned on the fan blade, 0 represents that the current pixel point is positioned outside the fan blade area,
Figure 320140DEST_PATH_IMAGE003
，
Figure 885113DEST_PATH_IMAGE004
for the row and column index numbers of the sliding window of the image respectively,
Figure 331138DEST_PATH_IMAGE012
the Tenengrad gradient corresponding to the current pixel point,
Figure 932013DEST_PATH_IMAGE013
is a preset threshold of Tenengrad gradient. In this example, the threshold is 0.01, and the determination result is shown in fig. 4;
8) setting the width of the sliding window to be w2, forming a sliding serial port with the window size of w2 xw 2, wherein the window width is 3 in the example;
9) as a typical sparsity evaluation index, the Hoyer statistical value is very sensitive to the sample distribution condition, and the Hoyer statistical value is adopted to evaluate the distortion degree of the sliding window. Moving the sliding window to the upper left corner of the gray image, and calculating the Hoyer statistical value (the distortion measurement value, reflecting the size of sparsity) of the elements in the current window, wherein the Hoyer statistical value algorithm is as follows:
Figure DEST_PATH_IMAGE032
wherein
Figure 8553DEST_PATH_IMAGE015
For the Hoyer statistic for the current sliding window,
Figure 60823DEST_PATH_IMAGE003
，
Figure 44960DEST_PATH_IMAGE004
for the row and column index numbers of the sliding window of the image respectively,
Figure 748342DEST_PATH_IMAGE016
is the size of the window or windows,
Figure 526942DEST_PATH_IMAGE033
is a point
Figure 66508DEST_PATH_IMAGE018
The gray scale information of (1).
10) Moving the sliding window to the right by one pixel, and calculating the Hoyer statistical value of the element in the current window;
11) continuously traversing all pixels of the current picture to form a Hoyer statistical value matrix, as shown in figure 5;
12) reserving the Hoyer statistical value of the fan blade area and setting the Hoyer statistical value of other areas to zero, wherein the specific algorithm for reserving the Hoyer statistical value is as follows:
Figure DEST_PATH_IMAGE034
；
wherein
Figure 306865DEST_PATH_IMAGE035
For the kept Hoyer statistics,
Figure 881066DEST_PATH_IMAGE021
is a matrix of the Hoyer statistical values,
Figure DEST_PATH_IMAGE036
is the judgment result of the pixel points in the fan blade area,
Figure 33830DEST_PATH_IMAGE003
，
Figure 44380DEST_PATH_IMAGE004
are the row-column index numbers of the image sliding window respectively.
13) Setting a Hoyer statistical value threshold, if the Hoyer statistical value is greater than the threshold, the pixel point is a defect characteristic, and the algorithm for judging the defect characteristic of the fan blade is as follows:
Figure 635899DEST_PATH_IMAGE037
；
wherein the content of the first and second substances,
Figure 799027DEST_PATH_IMAGE024
the defect judgment result of the current pixel point is that 1 indicates that the current pixel point is a defect position, 0 indicates that the current pixel point does not contain defect information,
Figure 919429DEST_PATH_IMAGE003
，
Figure 682855DEST_PATH_IMAGE004
for the row and column index numbers of the sliding window of the image respectively,
Figure DEST_PATH_IMAGE038
the Hoyer statistic value after being reserved for the current pixel point,
Figure 15747DEST_PATH_IMAGE039
is a preset Hoyer statistical value threshold value. In this example, the threshold is 0.01, and the specific result is shown in FIG. 6, which is clear from FIG. 6The fan blade defect characteristics are apparent.
Example 2: the invention relates to a method for detecting defects of an airborne fan blade, which is further explained by combining specific cases, and a flow chart of the method is shown as a figure 1, and comprises the following steps,
1) fixing a shooting device on the unmanned aerial vehicle, and shooting the fan blade to obtain a surface image of the fan blade, as shown in the attached figure 7;
2) converting the blade surface image to a grayscale image;
3) setting the sliding window width to w1, the window size is w1 × w1, in this example, the window width is 21;
4) moving the sliding window to the upper left corner of the gray image, and calculating the Tenengrad gradient of the elements in the current window, wherein the algorithm of the Tenengrad gradient is as follows:
Figure DEST_PATH_IMAGE040
wherein
Figure 485912DEST_PATH_IMAGE028
As the Tenengrad gradient information for the current sliding window,
Figure 42795DEST_PATH_IMAGE003
，
Figure 778670DEST_PATH_IMAGE004
for the row and column index numbers of the sliding window of the image respectively,
Figure 711991DEST_PATH_IMAGE005
is the size of the window or windows,
Figure 833399DEST_PATH_IMAGE006
is a point
Figure 561184DEST_PATH_IMAGE007
The information of the gradient in the x-direction,
Figure 784355DEST_PATH_IMAGE029
is a point
Figure 990208DEST_PATH_IMAGE030
Gradient information along the y-direction.
5) Moving the sliding window to the right by one pixel, and calculating the Tenengrad gradient of the element in the current window;
6) continuously traversing all pixels of the current picture to form a Tenengrad gradient matrix, as shown in figure 8;
7) setting a Tenengrad gradient threshold, if the Tenengrad gradient is greater than the threshold, taking the pixel as a fan blade area, and adopting a fan blade area judgment algorithm as follows:
Figure 231703DEST_PATH_IMAGE041
；
wherein
Figure 395968DEST_PATH_IMAGE011
The judgment result of the current pixel point is that 1 represents that the current pixel point is positioned on the fan blade, 0 represents that the current pixel point is positioned outside the fan blade area,
Figure 840856DEST_PATH_IMAGE003
，
Figure 850400DEST_PATH_IMAGE004
for the row and column index numbers of the sliding window of the image respectively,
Figure 228292DEST_PATH_IMAGE012
the Tenengrad gradient corresponding to the current pixel point,
Figure 287427DEST_PATH_IMAGE013
is a preset threshold of Tenengrad gradient. In this example, the threshold is 0.01, and the determination result is shown in FIG. 9;
8) setting the sliding window width to w2, the window size is w2 × w2, in this example, the window width is 3;
9) as a typical sparsity evaluation index, the Hoyer statistical value is very sensitive to the sample distribution condition, and the Hoyer statistical value is adopted to evaluate the distortion degree of the sliding window. Moving the sliding window to the upper left corner of the gray image, and calculating the Hoyer statistical value of the element in the current window, wherein the Hoyer statistical value algorithm is as follows:
Figure 485190DEST_PATH_IMAGE032
wherein
Figure 32846DEST_PATH_IMAGE015
For the Hoyer statistic for the current sliding window,
Figure 999665DEST_PATH_IMAGE003
，
Figure 240153DEST_PATH_IMAGE004
for the row and column index numbers of the sliding window of the image respectively,
Figure 174480DEST_PATH_IMAGE016
is the size of the window or windows,
Figure 525827DEST_PATH_IMAGE033
is a point
Figure 347152DEST_PATH_IMAGE018
The gray scale information of (1).
10) Moving the sliding window to the right by one pixel, and calculating the Hoyer statistical value of the element in the current window;
11) continuously traversing all pixels of the current picture to form a Hoyer statistical value matrix, as shown in figure 10;
12) reserving the Hoyer statistical value of the fan blade area and setting the Hoyer statistical value of other areas to zero, wherein the specific algorithm for reserving the Hoyer statistical value is as follows:
Figure DEST_PATH_IMAGE042
wherein
Figure 742231DEST_PATH_IMAGE035
For the kept Hoyer statistics,
Figure 649007DEST_PATH_IMAGE021
is a matrix of the Hoyer statistical values,
Figure 538465DEST_PATH_IMAGE036
is the judgment result of the pixel points in the fan blade area,
Figure 479876DEST_PATH_IMAGE003
，
Figure 45856DEST_PATH_IMAGE004
are the row-column index numbers of the image sliding window respectively.
13) Setting a Hoyer statistical value threshold, if the Hoyer statistical value is greater than the threshold, the pixel point is a defect characteristic, and the algorithm for judging the defect characteristic of the fan blade is as follows:
Figure 705507DEST_PATH_IMAGE043
；
wherein the content of the first and second substances,
Figure 398657DEST_PATH_IMAGE024
is the defect judgment result of the current pixel point, 1 represents that the current pixel point is the defect position, 0 represents that the current pixel point does not contain defect information,
Figure 194575DEST_PATH_IMAGE003
，
Figure 947767DEST_PATH_IMAGE004
for the row and column index numbers of the sliding window of the image respectively,
Figure 78403DEST_PATH_IMAGE038
the Hoyer statistic value after being reserved for the current pixel point,
Figure 575243DEST_PATH_IMAGE039
is a preset Hoyer statistical value threshold value. The threshold value is 0.003 in the present example, and the specific result is shown in fig. 11, and the defect characteristics of the fan blade can be obviously seen from fig. 11.
In conclusion, the method directly identifies the defects of the fan blade from the Tenengrad gradient and the Hoyer statistical value, has the characteristic of no influence of training data sample distribution, and can realize reliable and efficient state monitoring and defect identification of the fan blade.

Claims (7)

1. A method for detecting defects of blades of an airborne fan is characterized by comprising the following steps: the method comprises the following steps:
s1: acquiring a surface image of a fan blade, and converting the surface image of the fan blade into a gray image;
s2: setting a sliding window with the window size of w1 multiplied by w1, traversing all pixels of the current gray picture by using the sliding window, and calculating the Tenengrad gradient of the pixels in the current window to form a Tenengrad gradient matrix;
s3: setting a Tenengrad gradient threshold, and if the Tenengrad gradient is greater than the threshold, setting the pixel point as a fan blade area;
s4: setting a sliding window with the window size of w2 multiplied by w2, traversing all pixels of the current gray picture by using the sliding window, and calculating the Hoyer statistical value of the pixels in the current window until all pixels of the current picture are traversed to form a Hoyer statistical value matrix;
s5: keeping the Hoyer statistical value of the fan blade area in the step S4 and setting the Hoyer statistical values of other areas to zero;
s6: and setting a Hoyer statistical value threshold, and if the Hoyer statistical value of the fan blade area is greater than the threshold, determining that the pixel point is a defect characteristic.
2. The method for detecting the defects of the blades of the airborne fan according to claim 1, wherein the method comprises the following steps: the algorithm of the Tenengrad gradient is as follows:
Figure 624554DEST_PATH_IMAGE001
wherein
Figure DEST_PATH_IMAGE002
As the Tenengrad gradient information for the current sliding window,
Figure 479377DEST_PATH_IMAGE003
，
Figure DEST_PATH_IMAGE004
for the row and column index numbers of the sliding window of the image respectively,
Figure 446065DEST_PATH_IMAGE005
is the size of the window or windows,
Figure DEST_PATH_IMAGE006
is a point
Figure 44537DEST_PATH_IMAGE007
The information of the gradient in the x-direction,
Figure DEST_PATH_IMAGE008
is a point
Figure DEST_PATH_IMAGE009
Gradient information along the y-direction.
3. The method for detecting the defects of the blades of the airborne fan according to claim 1, wherein the method comprises the following steps: the judgment algorithm of the fan blade area is as follows:
Figure DEST_PATH_IMAGE010
；
wherein
Figure 983543DEST_PATH_IMAGE011
The judgment result of the current pixel point is 1The front pixel point is positioned on the fan blade, 0 represents that the current pixel point is positioned outside the fan blade area,
Figure 540426DEST_PATH_IMAGE003
，
Figure 525569DEST_PATH_IMAGE004
for the row and column index numbers of the sliding window of the image respectively,
Figure DEST_PATH_IMAGE012
the Tenengrad gradient corresponding to the current pixel point,
Figure 662152DEST_PATH_IMAGE013
is a preset threshold of Tenengrad gradient.
4. The method for detecting the defects of the blades of the airborne fan as claimed in claim 1, wherein the Hoyer statistic value algorithm is as follows:
Figure DEST_PATH_IMAGE014
wherein
Figure 783560DEST_PATH_IMAGE015
For the Hoyer statistic for the current sliding window,
Figure 245766DEST_PATH_IMAGE003
，
Figure 468937DEST_PATH_IMAGE004
for the row and column index numbers of the sliding window of the image respectively,
Figure DEST_PATH_IMAGE016
is the size of the window or windows,
Figure DEST_PATH_IMAGE017
is a point
Figure DEST_PATH_IMAGE018
The gray scale information of (1).
5. The method for detecting the defects of the blades of the airborne fan as claimed in claim 1, wherein in the step S5, the algorithm for retaining the Hoyer statistic is as follows:
Figure DEST_PATH_IMAGE019
wherein
Figure 65003DEST_PATH_IMAGE020
For the kept Hoyer statistics,
Figure DEST_PATH_IMAGE021
is a matrix of the Hoyer statistical values,
Figure 792917DEST_PATH_IMAGE022
is the judgment result of the pixel points in the fan blade area,
Figure 957182DEST_PATH_IMAGE003
，
Figure 667649DEST_PATH_IMAGE004
are the row-column index numbers of the image sliding window respectively.
6. The method for detecting the defects of the blades of the airborne fan according to claim 1, wherein the algorithm for judging the characteristics of the defects of the blades of the fan is as follows:
Figure DEST_PATH_IMAGE023
；
wherein the content of the first and second substances,
Figure 129724DEST_PATH_IMAGE024
is the defect judgment result of the current pixel point, 1 represents that the current pixel point is the defect position, 0 represents that the current pixel point does not contain defect information,
Figure 242036DEST_PATH_IMAGE003
，
Figure 46044DEST_PATH_IMAGE004
for the row and column index numbers of the sliding window of the image respectively,
Figure DEST_PATH_IMAGE025
the Hoyer statistic value after being reserved for the current pixel point,
Figure 961916DEST_PATH_IMAGE026
is a preset Hoyer statistical value threshold value.
7. The method for detecting the defects of the blades of the airborne fan according to claim 1, wherein the method comprises the following steps: and moving the sliding window to the upper left corner of the gray image, and sequentially moving one pixel from left to right and from top to bottom until the sliding window traverses all pixels of the current gray image.
CN202210613508.3A 2022-06-01 2022-06-01 Airborne fan blade defect detection method Pending CN114693684A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210613508.3A CN114693684A (en) 2022-06-01 2022-06-01 Airborne fan blade defect detection method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210613508.3A CN114693684A (en) 2022-06-01 2022-06-01 Airborne fan blade defect detection method

Publications (1)

Publication Number Publication Date
CN114693684A true CN114693684A (en) 2022-07-01

Family

ID=82130960

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210613508.3A Pending CN114693684A (en) 2022-06-01 2022-06-01 Airborne fan blade defect detection method

Country Status (1)

Country Link
CN (1) CN114693684A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117764992A (en) * 2024-02-22 2024-03-26 山东乔泰管业科技有限公司 Plastic pipe quality detection method based on image processing
CN117764992B (en) * 2024-02-22 2024-04-30 山东乔泰管业科技有限公司 Plastic pipe quality detection method based on image processing

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112200808A (en) * 2020-12-07 2021-01-08 领伟创新智能系统(浙江)有限公司 Strip steel surface defect detection method based on local Gini coefficient
CN112233067A (en) * 2020-09-21 2021-01-15 武汉钢铁有限公司 Hot rolled steel coil end face quality detection method and system
CN112368708A (en) * 2018-07-02 2021-02-12 斯托瓦斯医学研究所 Facial image recognition using pseudo-images
CN112819844A (en) * 2021-01-29 2021-05-18 山东建筑大学 Image edge detection method and device
CN114359274A (en) * 2022-03-16 2022-04-15 布鲁奇维尔通风设备启东有限公司 Ventilation equipment blade quality detection method, device and system based on image processing

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112368708A (en) * 2018-07-02 2021-02-12 斯托瓦斯医学研究所 Facial image recognition using pseudo-images
CN112233067A (en) * 2020-09-21 2021-01-15 武汉钢铁有限公司 Hot rolled steel coil end face quality detection method and system
CN112200808A (en) * 2020-12-07 2021-01-08 领伟创新智能系统(浙江)有限公司 Strip steel surface defect detection method based on local Gini coefficient
CN112819844A (en) * 2021-01-29 2021-05-18 山东建筑大学 Image edge detection method and device
CN114359274A (en) * 2022-03-16 2022-04-15 布鲁奇维尔通风设备启东有限公司 Ventilation equipment blade quality detection method, device and system based on image processing

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
吴鑫艺: "基于改进RFB和孪生网络的煤矿钻杆计数方法研究", 《中国优秀硕士论文全文库 工程科技Ⅰ辑》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117764992A (en) * 2024-02-22 2024-03-26 山东乔泰管业科技有限公司 Plastic pipe quality detection method based on image processing
CN117764992B (en) * 2024-02-22 2024-04-30 山东乔泰管业科技有限公司 Plastic pipe quality detection method based on image processing

Similar Documents

Publication Publication Date Title
Ginesu et al. Detection of foreign bodies in food by thermal image processing
CN109489724B (en) Tunnel train safe operation environment comprehensive detection device and detection method
CN111652098B (en) Product surface defect detection method and device
CN108827181B (en) Vision-based plate surface detection method
CN111060442B (en) Oil particle detection method based on image processing
CN110853018B (en) Computer vision-based vibration table fatigue crack online detection system and detection method
CN113870257B (en) Method and device for detecting and classifying defects of printed circuit board and computer storage medium
CN115100206B (en) Printing defect identification method for textile with periodic pattern
CN114881915A (en) Symmetry-based mobile phone glass cover plate window area defect detection method
CN112927223A (en) Glass curtain wall detection method based on infrared thermal imager
KR102242996B1 (en) Method for atypical defects detect in automobile injection products
CN112669286A (en) Infrared thermal image-based method for identifying defects and evaluating damage degree of external thermal insulation system of external wall
CN113034464B (en) Visual real-time detection method for defects of liquid crystal display under multiple backgrounds
CN113177925B (en) Method for nondestructive detection of fruit surface defects
CN114693684A (en) Airborne fan blade defect detection method
CN112634252A (en) Method for inspecting printed circuit
CN110887563B (en) Hyperspectral area array detector bad element detection method
CN111950351B (en) Agricultural machinery strain early diagnosis inspection system based on terahertz and visible light
CN111062939B (en) Method for rapidly screening quality of strip steel surface and automatically extracting defect characteristics
KR102114013B1 (en) Device and method for detecting defect of display
CN116523906B (en) Method and system for detecting optical performance of glass substrate
CN115615998B (en) Circular magnetic core side defect detection device and method
CN117495846B (en) Image detection method, device, electronic equipment and storage medium
CN117455917B (en) Establishment of false alarm library of etched lead frame and false alarm on-line judging and screening method
CN111091549B (en) Method for detecting breakage fault of crossed rod bodies of bottom parts of railway freight cars

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20220701

RJ01 Rejection of invention patent application after publication